Source – which-50.com
Artificial intelligence (AI) and machine learning are variously portrayed as the future of our workplaces or the death of employment as we know it. Somewhere between those two extremes lies the reality. These emerging technologies will be core parts not just of technology strategies, but of how we deliver services across our businesses.
We’re seeing it already in marketing through programmatic systems that determine the best places to make yourself be seen, in customer support through the use of chatbots and in security where vast volumes of data are scanned and analysed by computers that tell us where the real threats are.
IDC predicts that spending on machine learning and AI will leap from around $US12 billion last year to over $US57 billion over the next three years, with Deloitte adding that the number of AI and machine learning projects will double this year over last year and then double again by the end of this decade.
But what exactly is AI? Anthony Wong, a past president of the Australian Computer Society, recently spoke about AI. He said there are four types of artificial intelligence. The most basic form of AI is process robotics where systems mimic human action. We see this in factories, where a robot can perform an action that used to be completed by a human — but faster and with greater precision. We then move to intelligent automation, where systems mimic or augment human judgement. Cognitive automation systems augment human intelligence and, at the pinnacle, there’s artificial general intelligence — systems that can mimic human intelligence.
There’s not an expert on the planet who thinks we are particularly close to reaching artificial general intelligence. All the other levels in that model are founded upon two principles: the availability of data and algorithms that can make sense of the data and deliver a benefit.
Gartner Research Vice President and Distinguished Analyst Whit Andrews says that one of the mistakes many companies make when looking at the use of AI and machine learning is to try and do too much too early.
“Don’t fall into the trap of primarily seeking hard outcomes, such as direct financial gains, with AI projects,” said Andrews. “In general, it’s best to start AI projects with a small scope and aim for ‘soft’ outcomes, such as process improvements, customer satisfaction or financial benchmarking.”
Outside manufacturing environments, which have been using process robotics for decades, the best places to look for opportunities are where AI and machine learning can support and augment people — rather than replace them. He added, “Most transformational benefits of AI in the near term will arise from using it to enable employees to pursue higher-value activities”.
Further research by Andrews and his colleague Tom Austin suggests the best projects to focus on are ones where there’s a pressing business problem, rather than a massive redesign of a large-scale process. Small-scale adoption of AI is already happening through assistants that augment staff in tasks when they are using office productivity tools. But we’re also exposed to AI when we use services that analyse large swathes of data and produce meaningful insights — such as when we shop online.
Indeed this is becoming more important for retailers who operate in a world were consumers research extensively online but then carry the fruits of their effort into the store to complete the purchase. The data generated by that online activity provides the fuel that AI and machine learning can use to home in on the specific needs of each customer.
Real World Example
Fantastic Furniture is a case in point. A SAP C/4HANA customer that migrated from on-premise to cloud in 2016, Fantastic Furniture is already aware of the potential of machine learning to help it deliver better outcomes for customers.
Fantastic Furniture’s head of digital, Leigh McKnight, told CMO.com.au earlier this year that there is a lot of opportunity to help drive what customers see when they come to the company’s web site.
McKnight also told CMO that machine learning is used to drive product recommendations. “What we have on the mobile page for in-store shoppers is recommendations based on what merchandisers would recommend, plus what other customers purchased with that furniture item based on data from the web site.”
Muralidharan Surendran, a machine learning Innovation Architect from SAP, said businesses should target “low-hanging fruit”.
In his view, some of the areas businesses can look at are in human resources, customer relationship management, the legal area or activities where there is a lot of data, and business processes can be assisted by intelligent application of machine learning.
“The most critical part is the data set itself,” said Surendran. For example, he said, a data set like sales data requires some preparation. Information that is non-numerical, such as a product category, needs to be converted to a form that can be fed into machine learning systems, such as SAP Leonardo.
- Register your interest: Hear from senior executives including NAB’s Karen Ganschow, Iconic’s Anna Lee, Australia Post’s Andrew Walduck, Vectore’s Shamima Suntana, and Peoples Choice Credit Union’s Geoff Wenborn on the Real World Transformation panel and roundtable discussion in Sydney and Melbourne in late July and August. Register your interest today as places are limited.
When it comes to looking for partners to support your business on the AI and machine learning journey, Surendran said you need to find partners with experience who have invested in research and development that has been applied to deliver benefits to customers.
Surendran noted that many businesses try to do too much with machine learning and AI. But, he said, these tools don’t have to take over. He described a recent project he was involved with where the accuracy of an existing process was bolstered from 80 per cent to 85 per cent. That seemingly incremental change was enough to save the company millions of dollars and justify the project.
AI is no longer a technology of the future. It is here now and growing in its application and adoption. Over the next two to three years, the question won’t be whether or not to use AI and machine learning, but where and how to use it.